# Copyright (c) 2025, NVIDIA CORPORATION.  All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.


from typing import Callable, Optional

import lightning.pytorch as pl
import nemo_run as run
import torch
from lightning.pytorch.callbacks.callback import Callback
from megatron.core.distributed import DistributedDataParallelConfig

from nemo import lightning as nl
from nemo.collections.llm import Llama31Nemotron70BConfig, LlamaNemotronModel
from nemo.collections.llm.api import finetune, pretrain
from nemo.collections.llm.gpt.data.mock import MockDataModule
from nemo.collections.llm.gpt.data.packed_sequence import PackedSequenceSpecs
from nemo.collections.llm.peft import PEFT_STR2CLS
from nemo.collections.llm.recipes.finetune_default import default_finetune_recipe
from nemo.collections.llm.recipes.log.default import default_log, default_resume, tensorboard_logger
from nemo.collections.llm.recipes.optim.adam import distributed_fused_adam_with_cosine_annealing
from nemo.collections.llm.recipes.precision.mixed_precision import bf16_mixed
from nemo.collections.llm.recipes.tp_overlap_configs.userbuffers import (
    userbuffers_bf16_h100_h16384_tp8_cp2_mbs1_seqlen8192,
)
from nemo.lightning.pytorch.callbacks import GarbageCollectionCallback
from nemo.lightning.pytorch.callbacks.megatron_comm_overlap import MegatronCommOverlapCallback
from nemo.utils.exp_manager import TimingCallback

NAME = "llama31_nemotron_70b"


@run.cli.factory(name=NAME)
def model() -> run.Config[pl.LightningModule]:
    """
    Factory function to create a Llama-3.1-Nemotron-70B model configuration.
    This can be used for both Llama-3.1-Nemotron-70B-Instruct and Llama-3.1-Nemotron-70B-Reward.

    Returns:
        run.Config[pl.LightningModule]: Configuration for the Llama-3.1-Nemotron-70B model.

    Examples:
        CLI usage:
            $ nemo llm pretrain model=llama31_nemotron_70b ...

        Python API usage:
            >>> model_config = model()
            >>> print(model_config)
    """
    conf = run.Config(Llama31Nemotron70BConfig)
    conf.seq_length = 8192
    return run.Config(LlamaNemotronModel, config=conf)


def trainer(
    tensor_parallelism: int = 4,
    pipeline_parallelism: int = 4,
    pipeline_parallelism_type: Optional[torch.dtype] = torch.bfloat16,
    virtual_pipeline_parallelism: Optional[int] = 5,
    context_parallelism: int = 2,
    sequence_parallelism: bool = True,
    num_nodes: int = 4,
    num_gpus_per_node: int = 8,
    max_steps: int = 1168251,
    callbacks: Optional[list[run.Config[Callback]]] = None,
) -> run.Config[nl.Trainer]:
    """
    Configure the NeMo Lightning Trainer for Llama-3.1-Nemotron-70B model.

    This function sets up the distributed training strategy optimized for the large 70B model.

    Args:
        tensor_parallelism (int): Degree of tensor model parallelism.
        pipeline_parallelism (int): Degree of pipeline model parallelism.
        pipeline_parallelism_type (Optional[torch.dtype]): Data type for pipeline parallelism.
        virtual_pipeline_parallelism (Optional[int]): Size of virtual pipeline parallelism.
        context_parallelism (int): Degree of context parallelism.
        sequence_parallelism (bool): Whether to use sequence parallelism.
        num_nodes (int): Number of compute nodes to use.
        num_gpus_per_node (int): Number of GPUs per node.
        max_steps (int): Maximum number of training steps.
        callbacks (Optional[list[run.Config[Callback]]]): List of callback configurations.

    Returns:
        run.Config[nl.Trainer]: Configuration for the NeMo Lightning Trainer.

    Examples:
        CLI usage:
            $ nemo llm pretrain trainer=llama31_nemotron_70b ...

        Python API usage:
            >>> trainer_config = trainer(num_nodes=4, num_gpus_per_node=8)
            >>> print(trainer_config)

    Note:
        This configuration uses extensive parallelism to handle the large model size efficiently.
    """
    strategy = run.Config(
        nl.MegatronStrategy,
        tensor_model_parallel_size=tensor_parallelism,
        pipeline_model_parallel_size=pipeline_parallelism,
        pipeline_dtype=pipeline_parallelism_type,
        virtual_pipeline_model_parallel_size=virtual_pipeline_parallelism,
        context_parallel_size=context_parallelism,
        sequence_parallel=sequence_parallelism,
        gradient_as_bucket_view=True,
        ckpt_async_save=True,
        ckpt_parallel_load=True,
        ddp=run.Config(
            DistributedDataParallelConfig,
            check_for_nan_in_grad=True,
            grad_reduce_in_fp32=True,
            overlap_grad_reduce=True,
            overlap_param_gather=True,
            average_in_collective=True,
        ),
    )

    trainer = run.Config(
        nl.Trainer,
        accelerator="gpu",
        accumulate_grad_batches=1,
        callbacks=callbacks,
        devices=num_gpus_per_node,
        limit_test_batches=50,
        limit_val_batches=32,
        log_every_n_steps=10,
        max_steps=max_steps,
        num_nodes=num_nodes,
        plugins=bf16_mixed(),
        strategy=strategy,
        use_distributed_sampler=False,
        val_check_interval=2000,
    )

    return trainer


@run.cli.factory(target=pretrain, name=NAME)
def pretrain_recipe(
    dir: Optional[str] = None,
    name: str = "default",
    num_nodes: int = 1,
    num_gpus_per_node: int = 8,
    performance_mode: bool = False,
    fn: Callable = pretrain,
) -> run.Partial:
    """
    Create a pre-training recipe for Llama-3.1-Nemotron-70B model.

    This function sets up a complete configuration for pre-training, including
    model, trainer, data, logging, optimization, and resumption settings.

    Args:
        dir (Optional[str]): Directory for saving logs and checkpoints.
        name (str): Name of the pre-training run.
        num_nodes (int): Number of compute nodes to use.
        num_gpus_per_node (int): Number of GPUs per node.
        performance_mode (bool): If true, enables optimizations for maximum performance.
        fn (Callable): The pre-training function to use.

    Returns:
        run.Partial: Partial configuration for pre-training.

    Examples:
        CLI usage:
            $ nemo llm pretrain --factory llama31_nemotron_70b
            $ nemo llm pretrain --factory "llama31_nemotron_70b(num_nodes=4, name='my_70b_pretrain')"

        Python API usage:
            >>> recipe = pretrain_recipe(name="llama31_nemotron_70b_pretrain", num_nodes=4)
            >>> print(recipe)

    Note:
        This recipe is optimized for the large 70B model and requires significant computational resources.
    """
    recipe = run.Partial(
        fn,
        model=model(),
        trainer=trainer(
            num_nodes=num_nodes,
            num_gpus_per_node=num_gpus_per_node,
            callbacks=[run.Config(TimingCallback)],
        ),
        data=run.Config(MockDataModule, seq_length=8192, global_batch_size=512, micro_batch_size=1),
        log=default_log(dir=dir, name=name, tensorboard_logger=tensorboard_logger(name=name)),
        optim=distributed_fused_adam_with_cosine_annealing(max_lr=3e-4),
        resume=default_resume(),
    )

    if performance_mode:
        recipe = pretrain_performance_optimizations(recipe)

    return recipe


def pretrain_performance_optimizations(recipe: run.Partial) -> run.Partial:
    """
    Create a performance-optimized pre-training recipe for Llama-3.1-Nemotron-70B model.

    This method enables performance optimizations that may not be suitable for all use cases.
    It builds upon the standard pre-training recipe and adds additional performance enhancements.

    Args:
        recipe (run.Partial): Base pre-train recipe to which performance optimizations will be added

    Returns:
        run.Partial: Partial configuration for performance-optimized pre-training.

    Note:
        Use this method with caution and only when you need maximum performance.
        It may not be suitable for all hardware configurations or use cases.
    """

    # 'overlap_param_gather_with_optimizer_step' and 'align_param_gather' params are set automatically
    # by MegatronCommOverlapCallback. They are added here for user's knowledge.
    # overlap_param_gather_with_optimizer_step- Overlap param all-gather of first bucket with optimizer step.
    # align_param_gather- If true, all PP stages launch param all-gathers simultaneously, else
    # each PP stage launches independently as needed.

    recipe.trainer.callbacks.append(
        run.Config(
            MegatronCommOverlapCallback,
            tp_comm_overlap=True,
            tp_comm_overlap_cfg=userbuffers_bf16_h100_h16384_tp8_cp2_mbs1_seqlen8192,
            defer_embedding_wgrad_compute=True,
            wgrad_deferral_limit=50,
            overlap_param_gather_with_optimizer_step=False,  # Currently disabled due to an issue with checkpointing
            align_param_gather=True,
        )
    )

    return recipe


@run.cli.factory(target=finetune, name=NAME)
def finetune_recipe(
    dir: Optional[str] = None,
    name: str = "default",
    num_nodes: int = None,
    num_gpus_per_node: int = 8,
    model_type: str = 'instruct',
    peft_scheme: Optional[str] = 'lora',
    seq_length: Optional[int] = None,
    packed_sequence: Optional[bool] = None,
    performance_mode: bool = False,
) -> run.Partial:
    """
    Create a fine-tuning recipe for Llama-3.1-Nemotron-70B model.
    This can be used for both Llama-3.1-Nemotron-70B-Instruct and Llama-3.1-Nemotron-70B-Reward.
    By default it fine-tuned on Llama-3.1-Nemotron-70B-Instruct model.

    This function sets up a complete configuration for fine-tuning, including
    model, trainer, data, logging, optimization, and resumption settings.
    The recipe uses LoRA (Low-Rank Adaptation) for efficient fine-tuning, unless peft_scheme is set to None.

    Args:
        dir (Optional[str]): Directory for saving logs and checkpoints.
        name (str): Name of the fine-tuning run.
        num_nodes (int): Number of compute nodes to use.
        num_gpus_per_node (int): Number of GPUs per node.
        model_type (str): Type of model to fine-tune. Either 'instruct' or 'reward'.
        peft_scheme (Optional[str]): Name of the peft scheme to use for fine-tuning.
            Allowed values: 'lora'/'dora'/'none'/None.
        seq_length (int): Maximum number of tokens per microbatch.
        packed_sequence (Optional[bool]): If true, fine-tuning sequences will be packed into batches up to the given
            maximum seq_length for better efficiency. By default, this value equals performance_mode.
        performance_mode (bool): If true, enables optimizations for maximum performance.

    Returns:
        run.Partial: Partial configuration for fine-tuning.

    Examples:
        CLI usage:
            $ nemo llm finetune --factory llama31_nemotron_70b
            $ nemo llm finetune --factory "llama31_nemotron_70b(num_nodes=4, name='my_70b_finetune')"

        Python API usage:
            >>> recipe = finetune_recipe(name="llama31_nemotron_70b_finetune", num_nodes=4)
            >>> print(recipe)

    Note:
        This recipe uses the SQuAD dataset for fine-tuning. Be aware that fine-tuning a 70B model
        requires substantial computational resources.
    """
    # Default to unpacked data in normal mode and packed data in performance mode
    # once packing recipe is well tested, change this default to true
    if packed_sequence is None:
        packed_sequence = performance_mode

    # For unpacked sequence, most samples in SQuAD dataset are shorter than 2K
    if seq_length is None:
        seq_length = 4096 if packed_sequence else 2048

    if num_nodes is None:
        if peft_scheme is None or peft_scheme.lower() == 'none':
            num_nodes = 4
        elif peft_scheme.lower() in ['lora', 'dora']:
            num_nodes = 1

    if model_type.lower() == 'instruct':
        resume_path = 'nvidia/Llama-3.1-Nemotron-70B-Instruct-HF'
    elif model_type.lower() == 'reward':
        resume_path = 'nvidia/Llama-3.1-Nemotron-70B-Reward-HF'
    else:
        raise ValueError(f'Unknown model_type {model_type}. Expected one of "instruct" or "reward".')

    recipe = default_finetune_recipe(model(), resume_path, dir, name, num_nodes, num_gpus_per_node, packed_sequence)
    if peft_scheme is None or peft_scheme.lower() == 'none':
        recipe.trainer.strategy.tensor_model_parallel_size = 8
        recipe.trainer.strategy.pipeline_model_parallel_size = 4
        recipe.optim.config.lr = 5e-6
    elif peft_scheme.lower() in ['lora', 'dora']:
        recipe.peft = run.Config(PEFT_STR2CLS[peft_scheme.lower()])
        recipe.peft.dim = 16
        recipe.peft.alpha = 32
        recipe.optim.config.use_distributed_optimizer = False

        # some settings currently do not function correctly with LoRA
        recipe.model.config.cross_entropy_loss_fusion = False

        recipe.trainer.strategy.tensor_model_parallel_size = 8
        recipe.optim.config.lr = 1e-4
    else:
        raise ValueError(f"Unrecognized peft scheme: {peft_scheme}")

    # Sequence length settings in the model and dataset must agree
    recipe.model.config.seq_length = seq_length
    recipe.data.seq_length = seq_length
    if packed_sequence:
        recipe.data.dataset_kwargs = {'pad_to_max_length': True}
        recipe.data.packed_sequence_specs = run.Config(PackedSequenceSpecs, packed_sequence_size=seq_length)

    if performance_mode:
        recipe = finetune_performance_optimizations(recipe, peft_scheme)

    return recipe


def finetune_performance_optimizations(
    recipe: run.Partial,
    peft_scheme: str,
) -> run.Partial:
    """
    Modify the given recipe to optimize settings for performance.

    This method enables performance optimizations that may not be suitable for all use cases.
    Intended to build upon the standard fine-tuning recipe.

    Args:
        recipe (run.Partial): Base fine-tuning recipe to which performance optimizations will be added
        peft_scheme (Optional[str]): Name of the peft scheme to use for fine-tuning.
            Allowed values: 'lora'/'dora'/'none'/None.

    Returns:
        run.Partial: Partial configuration for performance-optimized fine-tuning.

    Note:
        Use this method with caution and only when you need maximum performance.
        It may not be suitable for all hardware configurations or use cases.
    """

    if not hasattr(recipe.trainer, "callbacks"):
        recipe.trainer.callbacks = []

    if peft_scheme is None or peft_scheme.lower() == 'none':
        recipe.trainer.strategy.tensor_model_parallel_size = 4
        recipe.trainer.strategy.pipeline_model_parallel_size = 4
        recipe.trainer.strategy.virtual_pipeline_model_parallel_size = 5
        recipe.trainer.plugins.grad_reduce_in_fp32 = False
        recipe.trainer.strategy.ddp = run.Config(
            DistributedDataParallelConfig,
            check_for_nan_in_grad=True,
            grad_reduce_in_fp32=False,
            overlap_grad_reduce=True,
            overlap_param_gather=True,
            average_in_collective=True,
        )
        recipe.trainer.callbacks.append(
            run.Config(
                MegatronCommOverlapCallback,
                tp_comm_overlap=True,
                defer_embedding_wgrad_compute=True,
                wgrad_deferral_limit=22,
            )
        )
    else:
        recipe.trainer.strategy.tensor_model_parallel_size = 2
        recipe.trainer.strategy.pipeline_model_parallel_size = 4
        recipe.trainer.strategy.virtual_pipeline_model_parallel_size = 5
        recipe.peft.target_modules = ['linear_qkv']

    recipe.trainer.strategy.sequence_parallel = True

    recipe.trainer.callbacks.append(run.Config(TimingCallback))
    recipe.trainer.callbacks.append(
        run.Config(
            GarbageCollectionCallback,
            100,
            100,
        )
    )

    return recipe